Our Team

We're a passionate group of experts who have worked together on complex graph problems for years:

• 3 dedicated researchers with backgrounds in knowledge representation

• 1 backend and 1 data engineers with extensive graph database experience

• 1 data scientists who specialize in graph analytics

• 1 natural language specialist and 1 generative AI developer

• 1 graph scientist focused on optimization algorithms

• 1 graph network expert

Company Name (if applicable)

Project details

Have you ever struggled with understanding why was the answer generated as it is? Can you rely on the answer that was given by the generative AI system? We doubt you'd be excited to get the answer "42" without proper explanation and run business or strategy decisions without full trust from the system.

Anyone who's worked with graph-based AI systems has encountered the "black box" problem where systems provide answers without explaining their reasoning. We've experienced these challenges firsthand, and they've motivated us to develop this proposal. Our vision is to create open, extensible, and scalable tools that illuminate the inference paths within graph systems - making them more transparent and trustworthy for users across domains.

We're focusing on three key areas that we believe will make the most significant impact:

1. Advanced Inference Path Identification and Tracing

We're determined to overcome the limitations of current path identification methods by developing more efficient and accurate techniques for tracing reasoning in graph systems.

We're particularly excited about:

• Implementing specialized graph traversal algorithms (BFS, DFS, A* search) optimized for inference path identification

• Developing pattern matching systems that can efficiently locate specific structural configurations in knowledge graphs

• Creating logical inference mechanisms that can trace the application of specific rules in rule-based systems

• Exploring reinforcement learning approaches for identifying optimal reasoning pathways in complex knowledge graphs

• Trace where conclusions came from

• Trace where inference engines "got stuck" or skipped over valid logic

For larger knowledge graphs with intricate reasoning chains, we're adapting specialized subgraph extraction techniques to focus on the most relevant portions of the graph. Early results suggest this approach significantly improves both efficiency and clarity of explanations.

Additionally, we're investigating causal graph representations (DAGs) to better capture and visualize the directional flow of reasoning from premises to conclusions.

2. Natural Language Verbalization of Inference Paths

We're a bit frustrated by the limitations of current explanation generation systems. Our goal is to create sophisticated verbalization techniques that can transform structured graph paths into clear, contextually appropriate natural language explanations.

We're exploring:

• Template-based generation systems with enhanced flexibility for various relationship types

• Rule-based verbalization frameworks that ensure grammatical correctness and coherence

• Neural approaches leveraging state-of-the-art language models for more natural-sounding explanations

• Hybrid systems that combine the reliability of templates with the fluency of neural models

One of our most ambitious goals is to develop context-aware explanation generation that adapts the level of detail, terminology, and format to match the user's background and needs. We're also creating visualization tools that complement textual explanations with intuitive diagrams of inference paths.

3. Comprehensive Evaluation Framework for Explanation Quality

We're tired of inadequate evaluation metrics that don't capture real explanation effectiveness. Our team is designing a robust evaluation framework that assesses the aspects that truly matter: accuracy, understandability, and usefulness of generated explanations.

Our evaluation framework will measure:

• Explanation fidelity (how accurately the explanation reflects the actual inference process)

• Linguistic quality (fluency, grammatical correctness, and readability)

• User comprehension (how well users understand the system's reasoning after reading the explanation)

• Task utility (how the explanations improve user performance on specific tasks)

• Domain appropriateness (suitability of explanations for different domains and use cases)

Beyond simple metrics, we're focusing on user-centered evaluation methods that capture the real-world impact of explanations on trust, decision-making, and system adoption.

Our project naturally aligns with the Hyperon framework and existing graph database systems by providing essential tools that enhance transparency and user understanding. We're committed to exploring integration with the MORK system to ensure consistent knowledge graph reasoning and validation within the OpenCog ecosystem.

1. Dong, G., & Pei, J. (2014). Natural Language Generation from Graph Structures. International Journal of Semantic Computing.

2. The Alan Turing Institute. Knowledge graphs.

3. Hwang, K., et al. (2022). A Decade of Knowledge Graphs in Natural Language Processing: A Survey.

4. Yin, Y., et al. (2023). Knowledge Graphs and Natural-Language Explanation Generation.

5. PuppyGraph. What is a Knowledge Graph? A Comprehensive Guide.

6. Ngomo, A.-C. N., et al. (2022). Generating Explanations in Natural Language from Knowledge Graphs.

7. Ontotext. Inferencing — GraphDB documentation.

8. IBM. What is Natural Language Generation (NLG)?

9. Datavid. Knowledge graph visualization: A comprehensive guide.

10. Goertzel, Ben, Vitaly Bogdanov, Michael Duncan, Deborah Duong, Zarathustra Goertzel, Jan Horlings, Matthew Ikle et al. "Opencog hyperon: A framework for agi at the human level and beyond." arXiv preprint arXiv:2310.18318 (2023).

11. Gephi. The Open Graph Viz Platform.

12. Fujitsu. Explainable AI Through Combination of Deep Tensor and Knowledge Graph.

13. Diffbot. Grounded Natural Language Generation with Knowledge Graphs.

14. Statworx. Explainable AI in practice: Finding the right method to open the Black Box.

15. PMC. Knowledge-graph-based explainable AI: A systematic review.

16. ACL Anthology. (2023). Knowledge-Grounded Natural Language Recommendation Explanation.

17. Springer Nature. Knowledge Graphs and Their Applications in Drug Discovery.

18. XAIWC. (2025). Explainable retrieval and graph augmented generation.

19. PMC. Querying knowledge graphs in natural language.

Open Source Licensing

Background & Experience

This proposal is one of five parts of a unified toolkit to be developed in parallel across a team of 14-16 developers and scientists. We have already a case of successful grant completion for DeepFunding.

Our team includes current employees from Yandex and Intel, former Linkedin, 3 university lecturers, and 7 PhDs. We're proud of our 10+ presidential awards, several patents and 30+ publications, including multiple technical books from a well known publishing house.

If we are awarded funding for 4 out of 5 proposals, we are committed to developing the 5th one at no additional cost.

We believe it’s better to have one toolkit than a kit of tools, that need to be duct-taped together.

We aim to deliver a robust, extensible, modular, production-ready ecosystem that can evolve with future RFPs, enabling seamless adoption, innovation, and collaboration. This approach will maximize the utility of knowledge graph fundamentals and pull in other innovative features and technologies from DeepFunding.